Load links between MAG and ProQuest

Load the links between records in MAG and in ProQuest. The links are stored as a table in the database. The links can be either between PhD graduates and MAG authors, or PhD advisors and MAG authors.

get_links(conn, from, min_score = 0.7, ...)

Arguments

conn

An object of class SQLiteConnection to a sqlite database.

from

A string with options to be queried: "advisors" or "graduates".

min_score

Minimum score for links to accept. Numeric between 0 and 1.

...

Additional arguments to be passed to make_tbl_output. If not specified, a lazily evaluated table without limit is returned. Partially specified arguments are completed with dots_tbl_output.

Value

A query of linked goid-AuthorId.

Details

For advisor links, a high threshold for min_score (0.95 or higher) is recommended, and the function throws a message if that is not the case. The reason for this is that there seem to be many false positives when using a lower threshold.

Examples

conn <- connect_to_db(db_example("AcademicGraph.sqlite"))
#> The database connection is: 
#> src:  sqlite 3.38.5 [/tmp/RtmptPxKJr/temp_libpath72cc939d74f39/magutils/extdata/AcademicGraph.sqlite]
#> tbls: AuthorAffiliation, FieldsOfStudy, FirstNamesGender, author_coauthor,
#>   author_output, current_links, current_links_advisors, pq_advisors,
#>   pq_authors, pq_fields_mag, pq_unis
graduate_links <- get_links(conn, from = "graduates", min_score = 0.7)
advisor_links <- get_links(conn, from = "advisors", min_score = 0.7)
#> Note: At the moment, using a link score below 0.95 for advisors can result in suspiciously many false positives. Carefully inspect the linked records you use.